The majority of developments in cereal science in the recent past have concentrated primarily on the functionality of the gluten protein sub-units and their role in bakery systems. This has been greatly facilitated by the abundance of natural variation between cultivators for the gluten protein sub-unit components.
In contrast, although flour from commercially grown wheat varieties contains approximately 75–85% starch, the role of starch from a breeding perspective has been overlooked; this is largely due to the difficulty of measuring differences in starch structure. Of the limited amount of work that has been carried out however, there appears to be a lack of natural variation between different wheat cultivars. With the advent of recombinant DNA and gene transfer technologies it is now possible to create new variation in planta, therefore directly modifying starch composition in wheat becomes a realistic target.
Starch is the major form of carbon reserve in plants, constituting 50% or more of the dry weight of many storage organs, e.g. tubers, seeds of cereals. Starch is used in numerous food and industrial applications. In many cases, however, it is necessary to modify the native starches, via chemical or physical means, in order to produce distinct properties to suit particular applications. It would be highly desirable to be able to produce starches with the required properties directly in the plant, thereby removing the need for additional modification. To achieve this via genetic engineering requires knowledge of the metabolic pathway of starch biosynthesis. This includes characterization of genes and encoded gene products which catalyse the synthesis of starch. Knowledge about the regulation of starch biosynthesis raises the possibility of “re-programming” biosynthetic pathways to create starches with novel properties that could have new commercial applications.
The most significant property of starch derives from the ability of the native granular form to lose its order and to swell and absorb water upon suitable treatment, thereby conferring viscosity and texture, in a process known as gelatinisation. Gelatinisation has been defined (W A Atwell et al, 1988) as “. . . the collapse (disruption) of molecular orders within the starch granule manifested in irreversible changes in properties such as granular swelling, native crystallite melting, loss of birefringence, and starch solubilisation. The point of initial gelatinisation and the range over which it occurs is governed by starch concentration, method of observation, granule type, and heterogeneities within the granule population under observation”.
14 molecules of water per molecule of anhydrous glucose, i.e. a minimum of 75% water, are necessary for full starch gelatinisation (Donovan, 1979). Starch gelatinisation is usually caused by heat, but can be caused by physical damage and some chaotropic agents, mainly dimethylsulphoxide (DMSO), urea, calcium chloride, strong base and acid.
The various events taking place during gelatinisation can be followed by various methods, including birefringence, X-ray diffraction, differential scanning calorimetry (DSC), 13C NMR. Swelling can be monitored by various methods, particularly rheology.
Differential scanning calorimetry (DSC) is a destructive method which records an endothermic event on heating of granules, generally thought to measure the temperature and the endothermic energy (delta H) required for the melting of the native crystallites. Starch gelatinisation temperature is independent of water content above 75% water (described as excess water), but increases when water is limited (Donovan, 1979).
The rate and extent of starch granule swelling upon heating dictate the type of viscosity development of aqueous starch suspensions on heating. Swelling behaviour is therefore of utmost technological importance. Viscosity increase on heating can be conveniently measured by a Brabender amylograph (Brabender is a Trade Mark) (Kennedy and Cabalda, 1991) or using a Rapid Visco analyser (Rapid Visco is a Trade-Mark from Newport Scientific, Australia). FIG. 1 is a typical viscoamylgraph profile for wheat starch, produced in this way, showing changes in starch during and after cooking. As starch granules swell on uptake of water, in, a process known as pasting, their phase volume increases, causing an increase in viscosity. The onset of pasting is indicated at A in FIG. 1. Peak viscosity, indicated at B in FIG. 1, is achieved when maximum phase volume is reached. Shear will then disrupt/cause fragmentation of the swollen granules, causing the viscosity to decrease. Complete dispersion is indicated at C in FIG. 1. This has been confirmed by an oscillatory rheology study of starch pastes at various stages of the viscosity profile (Svegmark and Hermansson, 1990). The viscosity onset temperature and peak viscosity are indicative of the initiation and extent of swelling, respectively. On cooling, leached amylose forms a network in a process involving reassociation of molecules, or retrogradation, causing an increase in viscosity as indicated at D in FIG. 1. Retrogradation (or set-back) viscosity is therefore indicative of the amount of amylose leached out of the granules.
The properties of wheat starch are useful in a large number of applications and also non-food (paper, textiles, adhesives etc.) applications. However, for many applications, properties are not optimum and various chemical and physical modifications well known in the art are undertaken in order to improve useful properties. Two types of property manipulation which would be of use are: the controlled alteration of gelatinisation and pasting temperatures; and starches which suffer less granular fragmentation during pasting than conventional starches.
Currently the only ways of manipulating the gelatinisation and pasting temperatures of starch are by the inclusion of additives such as sugars, polyhydroxy compounds of salts or by extensive physical or chemical pre-treatments. The reduction of granule fragmentation during pasting can be achieved either by extensive physical pre-treatments or by chemical cross-linking. Such processes are inconvenient and inefficient. It is therefore desirable to obtain plants which produce starch which intrinsically possesses such advantageous properties.
Starch consists of two main glucose polysaccharides: amylose and amylopectin. Amylose is a generally linear polymer comprising α-1,4 linked glucose units, while amylopectin is a highly branched polymer consisting of an α-1,4 linked glucan backbone with α-1,6 linked glucan branches. In wheat endosperm amylopectin constitutes approximately 70% of the total starch content, with the balance being amylose. Amylopectin is synthesised through the concerted action of several enzymes, including soluble starch synthase(s) (SSS), starch branching enzyme(s) (SBE), starch de-branching enzyme(s) (DBE). The physical properties of starch are strongly affected by the relative abundance of amylose and amylopectin, therefore SSSs, SBEs and DBEs play a key role in determining both starch quantity and quality. As such, one approach to manipulating starch structure would be to modify the expression of the enzymes involved in starch biosynthesis in the endosperm using a transgenic approach.
SBE catalyses the formation of the α-1,6 linkages, creating branch points in the growing starch molecule, via hydrolysis of an α-1,4 linkage followed by reattachment of the released α-1,4-glucan chain to the same or another glucosyl chain. This reaction also provides a new non-reducing end for further elongation of the original α-1,4-glucan chain.
Multiple isoforms of starch branching enzyme have been described, biochemically, from a number of species including maize (Boyer and Preiss, 1978), rice (Nakamura et al., 1992), pea (Smith, 1988), potato (Khoshnoodi et al., 1993) and wheat (Morell et al., 1997). More recently, genomic and cDNA sequences for SBE have been characterised from several species including maize (Baba et al., 1991; Fisher et al., 1993; Gao et al. 1997) pea (Burton et al., 1995), potato (Kossmann et al., 1991), rice (Nakamura and Yamanouchi, 1992; Mizuno et al., 1993), Arabidopsis (Fisher et al., 1996), cassava (Salehuzzaman et al., 1992), and wheat (Rapellin et al., 1997, Nair et al., 1997, Rahman et al., 1997). Sequence alignment of these SBEs revealed a high degree of sequence conservation at the amino acid level and that the SBEs may be grouped into two distinct families, generally known as SBEI and SBEII. Further, analysis indicates that within a species there is generally of the order of 50% homology between the two families, SBEI and SBEII, while there is often greater homology within the two families between species.
Maize is unusual in that the maize SBEII family is thought to comprise two different members, known as SBEIIa and SBEIIb. There has been controversy over whether the SBEIIa and IIb enzymes are in fact a) encoded by genes at two different loci, and b) whether the genes represent different alleles at a single locus. Fisher et al (1996) and Gao et al (1997) have provided evidence that SBEIIa and SBEIIb are encoded by independent genes. However, there is no conclusive evidence that both isoforms exist together in any one maize genotype. The DNA clones for the two published gene sequences were purified from different genotypes of maize and it is thus possible that they represent different alleles of a single locus. In summary, in maze, three distinct SBE genes have been characterised to date (Baba et al., 1991; Fisher et al., 1993; Gao et al., 1997). SBEI is distinct from SBEIIa and SBEIIb in amino acid composition, substrate specificity, kinetic properties, and immunological reactivities, whereas SBEIIa and SBEIIb are similar in these respects (Guan and Preiss, 1993; Preiss 1991; Takeda et al., 1993). At the amino acid level the sequence exhibits approximately 50% homology with the SBEIIa and SBEIIb sequences, whereas SBEIIa and SBEIIb exhibit approximately 80% homology to each other.
Prior to the present invention, maize was unique in having SBEIIa- and SBEIIb-type enzymes. Although Arabidopsis has two SBEII family members, the subdivision in Arabidopsis does not appear to conform to that seen in maize: the Arabidopsis sub-family members do not obviously fall into the IIa and IIb categories as do the maize sequences. Both of the Arabidopsis SBEII genes have similar levels of homology to both the maize SBEII genes, SBEIIa and SBEIIb, but the similarities are not sufficient to be able to place the Arabidopsis genes into the same SBEIIa and SBEIIb categories as for maize. Indeed, the data, if anything, suggests that the Arabidopsis SBEII genes do nor fall into the maize IIa and IIb categories. For barley, two forms of SBEII had been partly characterised. Although these have been called SBEIIa and SBEIIb, only a very limited amount of sequence information had been published (Sun et al, 1995) and it was not possible to infer or conclude that these forms correspond to the IIa and IIb categories of maize. In fact, based on the available barley sequence information both of the barley SBEII sequences (SBEIIa and SBEIIb) would appear to show greater homology to maize SBEIIa than to maize SBEIIb.
For all other plant species for which SBEII sequences have been identified and published, including potato, pea, rice, cassava, wheat and barley, no subdivision of the SBEII family comparable to the SBEIIa and SBEIIb division of maize has been made.
Studies of purified SBEI and SBEII demonstrate that these isoforms differ in their specificity for a substrate with respect to both chain length and degree of branching. In maize, SBEI and SBEII show distinct branching activities in vitro, with SBEI showing a higher rate of branching of an amylose substrate when compared to SBEII whereas both SBEIIa and IIb show higher rates of branching than SBEI when acting upon an amylopectin substrate (Guan and Preiss, 1993). Furthermore, maize SBEI preferentially transfers longer glucan chains (average chain length=24) than SBEII (average chain length=21(IIa) and 22(IIb)) (Takeda et al., 1993). A similar observation has been reported for SBEI and SBEII isoforms from wheat and pea (Morell et al., 1997; Smith, 1988). Mutational studies in maize, rice and pea demonstrate that high amylose mutants in each case are deficient in the branching enzyme activity analogous to maize SBEII (Martin and Smith, 1995; Morell et al., 1995). However, the linkage between the biochemical observations and the genetic evidence suggesting the differences in the roles remains unclear.
The present invention is based on the unexpected discovery of a novel class of SBEII genes in wheat, referred to herein as SBEII-1. The novel SBEII-1 gene sequence has strong homology with the maize SBEIIb gene. The wheat SBEII-1 genes are thought to be functionally equivalent to the maize SBEIIb gene, and on this basis it is believed that manipulation of the wheat SBEII-1 gene is likely to influence starch properties including starch gelatinisation temperature, in a manner analogous to manipulation of the maize SBEIIb gene as described in WO 97/22703.
In summary, although two different SBEII gene sequences are known from maize, Arabidopsis and barley, as discussed above, prior to the present invention there was no reason to expect that wheat would show a similar sub-division of SBEII genes as is seen for maize. The two Arabidopsis SBEII genes show a different sub-division, and prior to the present invention there was insufficient evidence to determine whether the two barley SBEII sequences belonged to the maize-type sub-division. That is, prior to the present invention there was no reason to expect that wheat would have two similar SBEII members comparable to those of maize. Subsequent to the present invention Sun et al (1998) have presented data which indicates that the barley sequences do indeed sub-divide in a similar manner to the maize SBEIIa and IIb sequences and the wheat SBEII-2 and SBEII-1 sequences discussed in this document.
The present inventors have used the high degree of sequence conservation between several SBE gene sequences to design oligonucleotide primers to motifs which are specific to either SBEI or SBEII families and have used these primers to amplify cDNA sequences from developing endosperm of wheat.
When this work was started, a single partial length wheat SBE cDNA clone had been reported (Mousley, 1994). Multiple sequence alignment of this wheat SBE sequence with other published SBE sequences from a number of plant species revealed a number of motifs which were highly conserved. Oligonucleotide primers designed to be complementary to these motifs were used to clone 3′ partial length cDNA clones of wheat SBEII. Alignment of the cDNA clone sequences indicated that the clones could be divided into two classes, which the inventors have designated SBEII-1 and SBEII-2, which showed greater than 90% similarity to members within a class but only 60% similarity between classes. Significantly, comparison between representative sequences from each class with previously identified wheat SBEII clones, pWBE6 (Mousley, 1994) and SBEII (Nair et al., 1997), showed that each appear to be homologues of the SBEII-2 class. The cloning of a wheat SBEII-1 cDNA is novel.